ggml-org/llama.cpp

Pull Requests Commits

sycl: Improve mul_mat_id memory efficiency and add BF16 fast path (#22119)

qnixsynapse committed 193 days ago

Verified 225088ea

mtmd: also support LLAMA_ROPE_TYPE_NONE (#22242)

ngxson committed 193 days ago

Verified 82d3f4d3

server: ignore reasoning content from transcription api (#21905)

ngxson committed 193 days ago

Verified 17f62451

mtmd, llama : Update HunyuanVL vision-language model support (#22037)

ManaEstras committed 193 days ago

Verified 7bfe60fd

common: Refactoring sampler parameters (#20429) (#22233)

ezturner committed 193 days ago

Verified 750579ff

common/chat, server: refactor, move all conversion functions to common, add tests (#20690)

pwilkin committed 193 days ago

Verified 134d6e54

ggml-webgpu(shader): support conv2d kernels. (#21964)

Constannnnnt committed 193 days ago

Verified ca7f7b7b

hexagon: add support for FILL op (#22198)

aparmp-quic committed 194 days ago

Verified 0dedb9ef

ggml-webgpu: reset CPU/GPU profiling time when freeing context (#22050)

yomaytk committed 194 days ago

Verified 2799d933

server: allow cancel loading model (#21814)

ngxson committed 194 days ago

Verified 04fe84b6

Hexagon: DAIG op (#22195)

shreyajn committed 194 days ago

Verified 5a4cd674

hexagon: fix missing v79 entry in libggml-htp.inf (#22194)

mengshengwu committed 194 days ago

Verified 2248799a

spec : reset i_last when low acceptance streak occurs (#22168)

treo committed 194 days ago

Verified 72d693e4

mtmd: Add support for Reka Edge 2603 (#21616)

kwajiehao committed 194 days ago

Verified 98d2d288

arg : add --spec-default (#22223)

ggerganov committed 194 days ago

Verified 84652b80

openvino: driver setup, CI split, thread safety, and NPU optimizations (#21944)

wine99 committed 194 days ago

Verified 52f1096f

vendor : update cpp-httplib to 0.43.1 (#22143)

cabelo committed 194 days ago

Verified 606fa42f

metal : workaround macOS GPU interactivity watchdog (#22216)

ggerganov committed 194 days ago

Verified 7fc1c4ef

vulkan: Support F16 OP_FILL (#22177)

jeffbolznv committed 194 days ago

Verified 82209efb

mtmd: correct mtmd_decode_use_mrope() (#22188)

ngxson committed 194 days ago

Verified 9998d88b

llama-ext : fix exports (#22202)

ggerganov committed 194 days ago

Verified cd03ec76

ggerganov committed 194 days ago

4889afba

ggml : bump version to 0.10.0 (ggml/1463)

ggerganov committed 194 days ago

041fe83d

fit-params : refactor + add option to output estimated memory per device (#22171)

ggerganov committed 194 days ago

Verified cfe9838d

server : fix hardcoded proxy connection timeout in router mode (#18760) (#22003)

xris99 committed 194 days ago

Verified ff6b1062

ggml-cuda: flush legacy pool on OOM and retry (#22155)

leonardHONG committed 195 days ago

Verified 97895129

mtmd: correct get_n_pos / get_decoder_pos (#22175)

ngxson committed 195 days ago

Verified 86f8daac

server : remove /api endpoints (#22165)

ggerganov committed 195 days ago

Verified cf8b0dbd

Tensor-parallel: Fix delayed AllReduce on Gemma-4 MoE (#22129)

gaugarg-nv committed 195 days ago

Verified fd6ae4ca

TP: fix 0-sized tensor slices, AllReduce fallback (#21808)

JohannesGaessler committed 195 days ago

Verified fb19f94c

Newer Older